Re: virus: RE: A most unscientific poll
Wed, 6 Nov 1996 23:50:48 -0600 (CST)

On Tue, 5 Nov 1996, Lior Golgher wrote:

> I've read an article about that theory of correlation between the order
> of birth and rebellious nature, and it seemed to me as an ordinary
> emphirical claim. It didn't claim that your rebellious nature is
> determined by birth and there's nothing you can do about it, it simply
> examplified the idea of order of birth as *one of the factors* which may
> effect your future rebellious acts. Just to step another foot, my
> immidiate impression based mainly on social reasoning was that this
> theory IS logical.


If I run enough statistical tests...I'll have to consider testing whether
the test results are random!

Seriously. Over a sufficiently large sample which, in Objective Reality,
has ratio X on a true/false property [keep things simple; same thing
holds in more complex cases] consider tests to falsify this X as the
correct ratio: a 5% significant test will say X is excluded 5% to the
time; a 1% significant test will say X is excluded 1% of the time, and
so on. This is the definition of "significance of a statistical test".

So, with enough tests, we can actually start testing the tests for
randomness [or lack thereof.]

/ Towards the conversion of data into information....
/ Kenneth Boyd